Grok 4.5 API Pricing
Grok 4.5 is SpaceXSI's previous-generation frontier model for coding, agentic tasks, and knowledge work, superseded by Grok 4.6 on August 12, 2026. The official SpaceXSI (formerly xAI) pricing page lists $2.00/M input, $0.30/M cached input, and $6.00/M output; the model page lists 500K context and text + image input. Cached input was cut from $0.50/M to $0.30/M on August 13, 2026 — the newer Grok 4.6 launched at the $0.50/M cache price this tier used to carry, which leaves Grok 4.5 the cheaper of the two on cache-heavy work.
Grok 4.5 API cost calculator.
Live calculator pre-loaded with Grok 4.5's sub-200K rates from SpaceXSI's pricing page. The model supports configurable reasoning, tool use, code execution, and text plus image input. Send a prompt of 200K tokens or more and every rate below doubles, across the whole request.
Grok 4.5 cost on real workloads.
Tool-calling agent
Research pack synthesis
Power-user chat
Knowledge base RAG
This is a chars-per-token approximation, not a real tokenizer. Actual tokens vary by language, code density, and tool-call overhead — counts are typically ±10–20% off for English prose, more for code or non-Latin scripts. For exact billing, use the vendor's official tokenizer.
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| Grok 4.5 Current | $2.00 cache $0.30 | $6.00 | $1.04 agentic 92/8 | 500K | Coding, agents, and knowledge work |
| Grok 4.7 | $2.00 cache $0.50 | $6.00 | $1.19 +15% | 500K | Newest frontier Grok, dearer cache |
| Grok 4.6 | $2.00 cache $0.50 | $6.00 | $1.19 +15% | 500K | Next frontier Grok, dearer cache |
| Grok 4.3 | $1.25 cache $0.20 | $2.50 | $0.558 −46% | 1M | Lower-cost Grok chat and tools |
| Grok Build 0.1 | $1.00 cache $0.20 | $2.00 | $0.476 −2.2× | 256K | Dedicated coding workflows |
| Grok 4.20 Multi-Agent (0309) | $1.25 cache $0.20 | $2.50 | $0.558 −46% | 1M | Pinned long-context agent loops |
| Claude Sonnet 5.5 | $2.00 cache $0.10 | $10.00 | $1.29 +24% | 1M | Claude-class coding agents |
| GPT-5.6 Sol | $4.00 cache $0.40 | $20.00 | $2.73 +2.6× · promo | 1.05M | OpenAI frontier workloads |
| DeepSeek V4 Pro | $1.32 cache $0.044 | $3.96 | $0.569 −45% | 1M | Low-cost reasoning alternative |
Grok 4.5 has published scores on 9 benchmarks we track; its best placing is #41 of 99 on ARC-AGI-1. 4 of these scores come from an independent leaderboard; the rest are the lab's own reports.
| Benchmark | Score | Rank | Leader | Source |
|---|---|---|---|---|
| ARC-AGI-1 Reasoning | 87.2% ARC Prize verified, best result | #41 of 99 | 6 tied · 98.5% | Official leaderboard ARC Prize Foundation |
| ARC-AGI-2 Reasoning | 52.6% ARC Prize verified, best result | #44 of 105 | GPT-6 Astra · 95% | Official leaderboard ARC Prize Foundation |
| Terminal-Bench 3.0 Agents and computer use | 15.7% high effort | #4 of 9 | DeepSeek V4.1 Flash · 30% | Vendor-reported SpaceXSI (xAI) |
| APEX-Agents Professional work | 47.1% high effort | #3 of 6 | Kimi K3.1 Pro · 69.4% | Vendor-reported SpaceXSI (xAI) |
| ARC-AGI-3 Reasoning · Standard harness | 0.3% ARC Prize verified, best result | #17 of 24 | GPT-6 Astra · 62.7% | Official leaderboard ARC Prize Foundation |
| Harvey Legal Agent Benchmark Professional work | 12.9% high effort, run by Vals | #5 of 7 | Kimi K3.1 Pro · 23.7% | Vendor-reported SpaceXSI (xAI) |
| DeepSWE Coding | 54% high effort | #22 of 27 | Gemini 4 Argon · 77.9% | Vendor-reported SpaceXSI (xAI) |
| Terminal-Bench 4.0 Agents and computer use | 12.4% high effort, Grok Build harness, 330 trials, ±2.6 pts (95% CI) | #29 of 34 | Kimi K3.1 Pro · 65.7% | Official leaderboard Terminal-Bench (Stanford, Laude Institute) |
| FrontierCode Coding · v1.1 Extended (all 150 tasks) | 56.6% high effort | #4 of 4 | GPT-6 Astra · 64.5% | Vendor-reported SpaceXSI (xAI) |
Grok 4.5 API pricing FAQ.
Grok 4.5 pricing questions, with SpaceXSI token rates separated from tool-call and cache assumptions.
Q · 01 What is Grok 4.5's API price? +
grok-4.5 at $2.00/M input, $0.30/M cached input, and $6.00/M output. These are USD prices per 1M tokens for prompts below 200K tokens; at or above that threshold every rate doubles to $4.00 / $0.60 / $12.00.Q · 02 Does Grok 4.5 support prompt caching? +
$0.30/M, versus $2.00/M for fresh input — an 85% discount, and cheaper than the $0.50/M the newer Grok 4.6 charges. SpaceXSI recommends setting a prompt cache key so long conversations route to the same server and reliably hit cache.Q · 03 What context window does Grok 4.5 have? +
500,000 tokens for Grok 4.5, the same window as the newer Grok 4.6. Grok 4.3 and the Grok 4.20 variants are listed at 1M, so the 500K models are not the longest-context Groks.Q · 04 Which regions can I call Grok 4.5 from? +
us-east-1 and us-west-2 — US regions only. At launch the docs carried a note that the model was not yet in the API console for EU users; that note is no longer on the page, and no EU region has replaced it.Q · 05 Is Grok 4.5 only for coding? +
Q · 06 How does Grok 4.5 compare with Grok 4.3 on price? +
$2/$6 versus Grok 4.3's $1.25/$2.50. It is the newer model, but Grok 4.3 remains a cheaper option with a larger 1M context window.